core idea extraction system

a systematized way to simplify and extract core ideas, generalized across: latent vectorization of words/sketches/EEG, symbolic regression (learning the rule), extracting ideas from transcripts, extracting principles from examples, taking EEG and extracting essential features, generalized interpretable ICA/PCA, history/english argument formulation from sources, and running research experiments.

this is a meta-tool for abstraction itself. connects to the writing tools for the language side and to the AGENTS.md research for the prompt optimization application.

timeline

• [2026-03-28] captured — generalized abstraction across modalities